English
Related papers

Related papers: Past- and Future-Informed KV Cache Policy with Sal…

200 papers

Transformer-based large language models (LLMs) use the key-value (KV) cache to significantly accelerate inference by storing the key and value embeddings of past tokens. However, this cache consumes significant GPU memory. In this work, we…

Multimodal Large Language Models (MLLMs) have advanced unified reasoning over text, images, and videos, but their inference is hindered by the rapid growth of key-value (KV) caches. Each visual input expands into thousands of tokens,…

Artificial Intelligence · Computer Science 2026-04-08 Bowen Zeng , Feiyang Ren , Jun Zhang , Xiaoling Gu , Ke Chen , Lidan Shou , Huan Li

Autoregressive language models rely on a Key-Value (KV) Cache, which avoids re-computing past hidden states during generation, making it faster. As model sizes and context lengths grow, the KV Cache becomes a significant memory bottleneck,…

Computation and Language · Computer Science 2025-03-05 Nathan Godey , Alessio Devoto , Yu Zhao , Simone Scardapane , Pasquale Minervini , Éric de la Clergerie , Benoît Sagot

Long-context LLM inference is bottlenecked by the quadratic attention complexity and linear KV cache growth. Prior approaches mitigate this via post-hoc selection or eviction but overlook the root inefficiency: indiscriminate writing to…

Machine Learning · Computer Science 2026-01-29 Yen-Chieh Huang , Pi-Cheng Hsiu , Rui Fang , Ming-Syan Chen

Scaling Diffusion Transformers to generate high-resolution, long videos is constrained by the quadratic cost of self-attention, and existing sparse attention methods degrade under high sparsity. We show empirically that generation quality…

Computer Vision and Pattern Recognition · Computer Science 2026-05-29 Shihao Han , Hao Yang , Xinting Hu , Xiaofeng Mei , Yi Jiang , Xiaojuan Qi

Autoregressive video diffusion models have enabled real-time, action-conditioned world generation. However, sustaining a persistent world, where revisiting a previously seen viewpoint yields consistent content, remains an open problem. Full…

Computer Vision and Pattern Recognition · Computer Science 2026-05-22 Jung Yi , Minjae Kim , Paul Hyunbin Cho , Wooseok Jang , Sangdoo Yun , Seungryong Kim

We introduce GoldFinch, a hybrid Linear Attention/Transformer sequence model that uses a new technique to efficiently generate a highly compressed and reusable KV-Cache in linear time and space with respect to sequence length. GoldFinch…

Computation and Language · Computer Science 2024-07-18 Daniel Goldstein , Fares Obeid , Eric Alcaide , Guangyu Song , Eugene Cheah

Inference with transformer-based language models begins with a prompt processing step. In this step, the model generates the first output token and stores the KV cache needed for future generation steps. This prompt processing step can be…

Computation and Language · Computer Science 2024-10-14 Maxwell Horton , Qingqing Cao , Chenfan Sun , Yanzi Jin , Sachin Mehta , Mohammad Rastegari , Moin Nabi

KV cache restoration has emerged as a dominant bottleneck in serving long-context LLM workloads, including multi-turn conversations, retrieval-augmented generation, and agentic pipelines. Existing approaches treat restoration as a…

Distributed, Parallel, and Cluster Computing · Computer Science 2026-04-29 Sean Nian , Jiahao Fang , Qilong Feng , Zhiyu Wu , Fan Lai

KV cache quantization reduces the memory cost of long-context LLM inference, but introduces approximation error that is typically validated only empirically. Existing systems rely on average-case robustness, with no mechanism to detect or…

Machine Learning · Computer Science 2026-05-21 Dean Calver

Large language models (LLMs) utilize key-value (KV) cache to store historical information during sequence processing. The size of KV cache grows linearly as the length of the sequence extends, which seriously affects memory usage and…

Computation and Language · Computer Science 2026-01-16 Yijun Liu , Yixuan Wang , Yuzhuang Xu , Shiyu Ji , Yang Xu , Qingfu Zhu , Wanxiang Che

Large Language Models (LLMs) have shown remarkable comprehension abilities but face challenges in GPU memory usage during inference, hindering their scalability for real-time applications like chatbots. To accelerate inference, we store…

Computation and Language · Computer Science 2024-06-06 Dongjie Yang , XiaoDong Han , Yan Gao , Yao Hu , Shilin Zhang , Hai Zhao

The Key-Value (KV) cache in generative large language models (LLMs) introduces substantial memory overhead. Existing works mitigate this burden by offloading or compressing the KV cache. However, loading the entire cache incurs significant…

Computation and Language · Computer Science 2025-05-28 Dingyu Yao , Bowen Shen , Zheng Lin , Wei Liu , Jian Luan , Bin Wang , Weiping Wang

This work studies how to adaptively recompute key-value (KV) caches for diffusion large language models (DLMs) to maximize prediction accuracy while minimizing decoding latency. Prior methods' decoders recompute QKV for all tokens at every…

Computation and Language · Computer Science 2025-12-30 Quan Nguyen-Tri , Mukul Ranjan , Zhiqiang Shen

Despite the significant success of large language models (LLMs), their extensive memory requirements pose challenges for deploying them in long-context token generation. The substantial memory footprint of LLM decoders arises from the…

Machine Learning · Computer Science 2024-02-12 Amir Zandieh , Insu Han , Vahab Mirrokni , Amin Karbasi

Retrieval-Augmented Generation (RAG) systems suffer from severe time-to-first-token (TTFT) bottlenecks due to long input sequences. Existing KV cache reuse methods face a fundamental trade-off: prefix caching requires identical prefixes…

Machine Learning · Computer Science 2026-05-22 Bin Yang , Qiuyu Leng , Jun Zeng , Zhenhua Wu

Large language models have revolutionized natural language processing but face significant challenges of high storage and runtime costs, due to the transformer architecture's reliance on self-attention, particularly the large KV cache for…

Computation and Language · Computer Science 2026-05-29 Yuan Feng , Junlin Lv , Haoyu Guo , Yukun Cao , S Kevin Zhou , Xike Xie

Advanced autoregressive (AR) video generation models have improved visual fidelity and interactivity, but the quadratic complexity of attention remains a primary bottleneck for efficient deployment. While existing sparse attention solutions…

Computer Vision and Pattern Recognition · Computer Science 2026-05-26 Chengtao Lv , Yumeng Shi , Yushi Huang , Ruihao Gong , Shen Ren , Wenya Wang

Key-value (KV) cache compression has emerged as a critical technique for reducing the memory and latency overhead of autoregressive language models during inference. Prior approaches predominantly rely on query-key attention scores to rank…

Computation and Language · Computer Science 2025-09-19 Ayan Sengupta , Siddhant Chaudhary , Tanmoy Chakraborty

Recent advances in video generation have been dominated by diffusion and flow-matching models, which produce high-quality results but remain computationally intensive and difficult to scale. In this work, we introduce VideoAR, the first…

Computer Vision and Pattern Recognition · Computer Science 2026-01-15 Longbin Ji , Xiaoxiong Liu , Junyuan Shang , Shuohuan Wang , Yu Sun , Hua Wu , Haifeng Wang